Download Microdata User Guide

Microdata User Guide
Canada Survey of Giving, Volunteering and
Participating
2004
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Table of Contents
1.0
Introduction
............................................................................................................................... 5
2.0
Background
............................................................................................................................... 7
3.0
Objectives
............................................................................................................................... 9
4.0
Concepts and definitions ............................................................................................................ 11
5.0
Survey methodology for the provincial component................................................................. 15
5.1
Population coverage ......................................................................................................... 15
5.2
Sample design .................................................................................................................. 15
5.2.1 Stratification ......................................................................................................... 15
5.2.2 Sample allocation................................................................................................. 15
5.3
Sample selection............................................................................................................... 16
5.4
Sample size by province ................................................................................................... 17
6.0
Survey methodology for the territorial (northern) component................................................ 19
6.1
Population coverage ......................................................................................................... 19
6.2
Sample design .................................................................................................................. 19
6.2.1 Sample rotation .................................................................................................... 19
6.2.2 Modifications to the Labour Force Survey design in the territories for the
Canada Survey of Giving, Volunteering and Participating................................... 19
6.3
Sample size.......................................................................................................................20
7.0
Data collection ............................................................................................................................. 21
7.1
Questionnaire design ........................................................................................................ 21
7.2
Supervision and quality control ......................................................................................... 21
7.3
Data collection methodology............................................................................................. 21
7.3.1 Provincial component........................................................................................... 21
7.3.2 Territorial component ........................................................................................... 22
7.4
Non-response.................................................................................................................... 22
8.0
Data processing ........................................................................................................................... 23
8.1
Data capture......................................................................................................................23
8.2
Editing ...............................................................................................................................23
8.3
Coding of open-ended questions ...................................................................................... 23
8.4
Imputation .........................................................................................................................24
8.5
Creation of derived variables ............................................................................................ 24
8.6
Weighting ..........................................................................................................................25
8.7
Suppression of confidential information ............................................................................ 25
9.0
Data quality
............................................................................................................................. 27
9.1
Response rates .................................................................................................................27
9.1.1 Response to the provincial component................................................................ 27
9.1.2 Response to the territorial component................................................................. 28
9.2
Survey errors.....................................................................................................................29
9.2.1 Data collection...................................................................................................... 29
9.2.2 Data processing ................................................................................................... 29
9.2.3 Non-response and imputation.............................................................................. 29
9.2.4 Measurement of sampling error ........................................................................... 31
Special Surveys Division
3
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
10.0
Guidelines for tabulation, analysis and release ....................................................................... 33
10.1
Rounding guidelines.......................................................................................................... 33
10.2
Sample weighting guidelines for tabulation ...................................................................... 33
10.3
Definitions of types of estimates: categorical and quantitative ......................................... 34
10.3.1 Categorical estimates .......................................................................................... 34
10.3.2 Quantitative estimates ......................................................................................... 34
10.3.3 Tabulation of categorical estimates ..................................................................... 35
10.3.4 Tabulation of quantitative estimates .................................................................... 35
10.4
Guidelines for statistical analysis ...................................................................................... 35
10.5
Coefficient of variation release guidelines ........................................................................ 36
10.6
Release cut-offs for the Canada Survey of Giving, Volunteering and Participating ......... 38
11.0
Approximate sampling variability tables ................................................................................... 39
11.1
How to use the coefficient of variation tables for categorical estimates ........................... 40
11.1.1 Examples of using the coefficient of variation tables for categorical estimates... 41
11.2
How to use the coefficient of variation tables to obtain confidence limits......................... 44
11.2.1 Example of using the coefficient of variation tables to obtain confidence limits .. 45
11.3
How to use the coefficient of variation tables to do a t-test .............................................. 46
11.3.1 Example of using the coefficient of variation tables to do a t-test........................ 46
11.4
Coefficients of variation for quantitative estimates ........................................................... 46
11.5
Coefficient of variation tables............................................................................................ 47
12.0
Weighting
............................................................................................................................. 49
12.1
Weighting for the provincial component............................................................................ 49
12.2
Weighting for the territorial component ............................................................................. 52
13.0
Questionnaires ............................................................................................................................. 55
14.0
Structure of the files .................................................................................................................... 57
15.0
Variable naming conventions ..................................................................................................... 59
16.0
Record layout with univariate frequencies................................................................................ 61
4
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
1.0
Introduction
The Canada Survey of Giving, Volunteering and Participating (CSGVP) is one component of the
Voluntary Sector Initiative, a collaborative program of the federal government and the voluntary sector.
The CSGVP was conducted by Statistics Canada in the 10 provinces from mid-September to December,
2004 and from the end of August to mid-November, 2004 in the three territories.
This manual has been produced to facilitate the manipulation of the microdata file of the survey results.
Any questions about the data set or its use should be directed to:
Statistics Canada
Client Services
Special Surveys Division
Telephone: 613-951-3321 or call toll-free 1-800-461-9050
Fax: 613-951-4527
E-mail: [email protected]
Special Surveys Division
5
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
2.0
Background
In the course of their busy lives and many commitments, millions of Canadians make a conscious effort to
contribute to others and their communities through charitable giving, volunteering their time to charitable
and non-profit organizations and by helping individual Canadians directly on their own. In 1997, the
National Survey of Giving, Volunteering and Participating (NSGVP) provided the first comprehensive look
at the contributions that Canadians made to one another through their gifts of time and money. The
NSGVP was developed through a unique partnership of federal government departments and non-profit
and voluntary organizations that included the Canadian Centre for Philanthropy (now operating under the
name of Imagine Canada), Canadian Heritage, Health Canada, Human Resources Development Canada,
Statistics Canada and Volunteer Canada. Using a similar framework, this survey was conducted again in
2000 as part of the federal government’s Voluntary Sector Initiative. In 2001, the federal government
provided funding to establish a permanent survey program at Statistics Canada on charitable giving,
volunteering and participating. The survey itself was renamed the Canada Survey of Giving, Volunteering
and Participating (CSGVP) to distinguish it from surveys in other countries.
The establishment of a permanent series of surveys provided an opportunity to review the design of the
survey instrument to ensure that it would provide the highest quality information on an ongoing basis.
Consultations were held with a variety of stakeholders from the charitable and non-profit sector,
government and the academic community to identify ways to improve the survey. In 2004, survey data
were collected in the North (Yukon, Northwest Territories and Nunavut) for the first time, where a
representative sample of 1,332 respondents aged 15 and older participated in the survey. The sample
size in the 10 provinces increased from 14,724 respondents in 2000 to 20,832 in 2004 improving the
ability to provide estimates both at the provincial level and in the larger urban areas. The questionnaire
was revised in a number of ways, based on experience gained from the earlier surveys. Some questions
were changed to improve their clarity for respondents. Other questions were added to collect new
information of interest. A number of questions were also dropped from the survey. Because the survey is
now being conducted on a permanent basis, it may be possible to cycle sets of questions in and out of
the survey.
The survey platform was also changed. The NSGVP was administered to a sub-sample of respondents
to Statistics Canada’s Labour Force Survey (LFS). Because of concerns about demands being placed on
LFS respondents, the provincial component of the 2004 CSGVP was conducted as a Random Digit
Dialling (RDD) survey, in which respondents were recruited specifically to participate in the CSGVP.
The 2004 CSGVP provides a new way of measuring giving, volunteering and participating. It replaces the
way these behaviours were measured in the 1997 and 2000 NSGVPs. Because of these changes, it is
not appropriate to compare the results from the 2004 CSGVP with the previous NSGVP surveys.
Special Surveys Division
7
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
3.0
Objectives
The objectives of the Canada Survey of Giving, Volunteering and Participating (CSGVP) are threefold:
1) to collect national data to fill a void of information about individual contributory behaviours
including volunteering, charitable giving and civic participation;
2) to provide reliable and timely data to the System of National Accounts; and
3) to inform both the public and voluntary sectors in policy and program decisions that relate to the
charitable and voluntary sector.
Special Surveys Division
9
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
4.0
Concepts and definitions
This chapter outlines concepts and definitions of interest to the users.
Donor
A donor is a person who made at least one donation of money to a charitable or other non-profit
organization in the 12-month reference period preceding the survey.
Financial donation
A financial donation is money given to a charitable or other non-profit organization during the 12-month
reference period preceding the survey. Money given to the same organization, on multiple occasions, in
response to the same solicitation method, constitutes only one donation. For example, all money
donated to a particular religious institution over the 12 months preceding the survey, through a collection
at the place of worship, would be considered to be a single donation.
In-kind donation
This is a non monetary donation made to a charitable or other non-profit organization. Examples include
donations of clothing or household items and donations of food.
Industry and Occupation
The 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP) provides industry and
occupation information for employed persons only (i.e., regarding the job which the individual occupied
the week preceding the interview). For industry, statistics have been provided based on both the 1997
and 2002 North American Industry Classification Systems (NAICS). For occupation, both the 1991
Standard Occupation Classification (SOC) and the 2001 National Occupation Classification – Statistics
(NOC-S) have been used.
Informal volunteer
The CSGVP defines an informal volunteer (or a direct helper) as a person who helped someone on their
own, that is, not through a group or organization, in the 12-month reference period preceding the survey.
This includes help given directly to friends, neighbours and relatives, but excludes help given to anyone
living in the household. Since these activities are not provided through the structure of an organization,
they are not included under the definition of volunteering.
Labour force status
Labour force status designates the status of the respondent vis-à-vis the labour market. For the 2004
CSGVP, estimates of labour force status refer to the survey population aged 15 to 75, as respondents
aged 76 and older were not asked the related series of questions.
The three categories of labour force status are “employed”, “unemployed” and “not in the labour force”.
For the purposes of the CSGVP, the three categories of labour force status are defined as follows:
Employed
Employed persons are those who, during the week preceding the interview
a) did any work 1 at all at a job or business; or
1
Work includes any work for pay or profit, that is, paid work in the context of an employer-employee
relationship, or self-employment. It also includes unpaid family work, which is defined as unpaid work
contributing directly to the operation of a farm, business or professional practice owned and operated by a
related member of the same household. Such activities may include keeping books, selling products,
waiting on tables, and so on. Tasks such as housework or maintenance of the home are not considered
unpaid family work.
Special Surveys Division
11
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
b) had a job but were not at work due to factors such as own illness or disability, personal or
family responsibilities, vacation, labour dispute or other reasons (excluding persons on
layoff or between casual jobs).
Unemployed
Unemployed persons are those who, during the week preceding the interview
a) were on temporary layoff (excluding full-time students); or
b) were without work, and had actively looked for work in the past four weeks, (excluding
full-time students and retired persons).
Not in the labour force
Persons not in the labour force are those who had not worked during the week preceding the
interview and
a) were permanently unable to work; or
b) were full-time students who had a job but were absent from work as a result of a layoff or
because they were between casual jobs; or
c) were full-time students or retired persons who did not have a job and had looked for
work; or
d) did not have a job and did not look for work.
Mandatory community service
This is unpaid help provided to a group or organization that was mandated, or required, by a school, an
employer, a charitable or non-profit organization, or some other authority. The 2004 CSGVP includes
mandatory service under the definition of volunteering.
Organization classification
Respondents were asked to provide information on the organizations for which they volunteered and to
which they made donations. Respondents were first asked to provide the name of the organization. A
pick-list including the most common organizations reported in the 1997 and 2000 surveys was used. If the
organization cited by the respondent was not on this pick-list, the respondent was then asked to provide
information about what this organization does. This information was then used to group organizations into
broad categories.
To classify these organizations, the International Classification of Nonprofit Organizations (ICNPO) 2 was
used. Although they are classified according to their primary area of activity, some organizations operate
in multiple areas. A major advantage of the ICNPO system is that it is used widely by other countries and
thus allows for international comparisons. It has also been devised specifically to reflect the range and
nature of activities typically undertaken in the non-profit and voluntary sector. The ICNPO system
developed by the Johns Hopkins Comparative Nonprofit Sector Project, and modified for use in Canada,
groups organizations into 15 Major Activity Groups, including a catch-all “Not Elsewhere Classified”
category. These 15 Major Activity Groups are further grouped into 12 categories.
The 15 categories are as follows:
1) Arts and culture: includes organizations and activities in general and specialized fields of arts and
culture, including media and communications; visual arts, architecture, ceramic art; performing
arts; historical, literacy and humanistic societies; museums; and zoos and aquariums.
2) Sports and recreation: includes organizations and activities in general and specialized fields of
sports and recreation. Two sub-groups of organizations are included in this group: (1) amateur
sports (including fitness and wellness centres); and (2) recreation and social clubs (including
service clubs).
2
The classification is based on L.M. Salamon and H.K. Anheier, 1997. Defining the Nonprofit Sector: A
Cross-national Analysis. Manchester University Press.
12
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
3) Education and research: includes organizations and activities administering, providing, promoting,
conducting, supporting and servicing education and research. Three sub-groups are contained in
this group: (1) primary and secondary education organizations; (2) organizations involved in other
education (i.e., adult/continuing education and vocational/technical schools); and (3)
organizations involved in research (i.e., medical research, science and technology, and social
sciences). Note that organizations devoted primarily to education and research in the area of
specific medical conditions (e.g., Heart and Stroke Foundation of Canada, Canadian Cancer
Society) are included under category 5, Health.
4) Universities and colleges: includes organizations and activities related to higher learning. This
includes universities, business management schools, law schools and medical schools.
5) Health: includes organizations that engage primarily out-patient health-related activities and
health support services. Two sub-groups are included in this category: (1) mental health
treatment and crisis intervention; and (2) other health services (including public health and
wellness education, out-patient health treatment, rehabilitative medical services, and emergency
medical services). Also included in this category are organizations devoted primarily to education,
research or support services in the area of specific medical conditions (e.g., Heart and Stroke
Foundation, Canadian Cancer Society) as well as organizations providing support to the
terminally ill (e.g., hospices and other types of palliative care).
6) Hospitals: includes organizations that engage primarily in in-patient health care. Two sub-groups
are included in this category: (1) hospitals and rehabilitation; and (2) nursing homes.
7) Social Services: includes organizations and institutions providing human and social services to a
community or target population. Three sub-groups are contained in this category: (1) social
services (including organizations providing services for children, youth, families, the handicapped
and the elderly, and self-help and other personal social services); (2) emergency and relief; and
(3) income support and maintenance.
8) Environment: includes organizations promoting and providing services in environmental
conservation, pollution control and prevention, environmental education and health, and animal
protection. Two sub-groups are included in this category: (1) environment; and (2) animal
protection.
9) Development and housing: includes organizations promoting programs and providing services to
help improve communities and promote the economic and social well-being of society. Three subgroups are included in this category: (1) economic, social and community development (including
community and neighbourhood organizations); (2) housing; and (3) employment and training.
10) Law, Advocacy and Politics: includes organizations and groups that work to protect and promote
civil and other rights, advocate the social and political interests of general or special
constituencies, offer legal services or that promote public safety. Three sub-groups are contained
in this category: (1) civic and advocacy organizations; (2) law and legal services; and (3) political
organizations.
11) Grant-making, fundraising and voluntarism promotion: includes philanthropic organizations and
organizations promoting charity and charitable activities including grant-making foundations,
voluntarism promotion and support, and fund-raising organizations.
12) International: includes organizations promoting cultural understanding between peoples of
various countries and historical backgrounds as well as those providing relief during emergencies
and promoting development and welfare abroad.
13) Religion: includes organizations promoting religious beliefs and administering religious services
and rituals (e.g., churches, mosques, synagogues, temples, shrines, seminaries, monasteries
Special Surveys Division
13
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
and similar religious institutions), in addition to related organizations and auxiliaries of such
organizations.
14) Business and professional associations, unions: includes organizations promoting, regulating and
safeguarding business, professional and labour interests.
15) Groups not elsewhere classified.
The correspondence between the 12 category classification and the 15 category classification is as
follows:
12 Category ICNPO
15 Category ICNPO
1)
Culture and recreation
1)
2)
Arts and culture
Sports and recreation
2)
Education and research
3)
4)
Education and research
Universities and colleges
3)
Health
5)
6)
Health
Hospitals
4)
Social services
7)
Social services
5)
Environment
8)
Environment
6)
Development and housing
9)
Development and housing
7)
Law, advocacy and politics
10) Law, advocacy and politics
8)
Philanthropic intermediaries and voluntarism
11) Grant-making, fundraising and voluntarism promotion
9)
International
12) International
10) Religion
13) Religion
11) Business and professional associations, unions
14) Business and professional associations, unions
12) Groups not elsewhere classified
15) Groups not elsewhere classified
Participant
The CSGVP defines a participant as a person who was a member of at least one group, organization or
association in the 12-month reference period preceding the survey. This includes professional
organizations or unions; service clubs or fraternal organizations; political groups; cultural, educational, or
hobby related organizations; sports or recreation organizations; religious organizations; seniors’ or youth
groups; support or self-help programs; environmental groups; and community or school related
associations.
Reference period
For most questions in the CSGVP questionnaire, the reference period was the 12 months preceding the
interview. For the provincial component, interviews were conducted from September 13th to December
19th, 2004. For the territorial or northern component, interviews took place from August 30th to November
15th, 2004.
Volunteer
This is a person who volunteered, that is, who performed a service without pay, on behalf of a charitable
or other non-profit organization, at least once in the 12 month reference period preceding the survey. This
includes any unpaid help provided to schools, religious organizations, sports or community associations.
14
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
5.0
Survey methodology for the provincial component
In the 10 provinces, the 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP) was
administered between September 13, 2004 and December 19, 2004 as a Random Digit Dialling (RDD)
survey, a technique whereby telephone numbers are generated randomly by computer. Interviews were
conducted by telephone.
5.1
Population coverage
The target population consisted of the population 15 years of age or older residing in Canada’s 10
provinces, with the exception of the institutionalized population.
The surveyed population excluded persons living in households without a land phone line, i.e.,
those living in households with no phone or with only cell phones were excluded. It is estimated
that in 2004 approximately 4.2% 3 of households in the 10 provinces had no land line telephone,
1.5% having no phone and 2.7% having cell phones only. It is important to realize that although
these persons were excluded from the population surveyed, the estimates were weighted to
account for them. The underlying assumption is that the people in these households have the
same characteristics and behaviours as those surveyed.
5.2
Sample design
5.2.1
Stratification
The sample for the provincial component of CSGVP is based on a stratified design
employing probability sampling. The stratification was done at the province / census
metropolitan area (CMA) level. Twenty-seven strata were formed. Each province was
divided into a number of CMA strata (ranging from zero in Prince Edward Island to four in
Ontario) and one additional residual “non-CMA” stratum comprising the remainder of the
province.
5.2.2
Sample allocation
The sample size was determined in order to be able to produce:
1) cross-sectional estimates for volunteers provincially and for the three largest
CMAs;
2) cross-sectional estimates for non-volunteers provincially and for the three largest
CMAs;
3) national cross-sectional estimates for immigrants; and
4) national longitudinal estimates for those who change volunteer status.
It was determined that 40,000 responses would be required to meet these objectives. A
power allocation (power = 0.20) was used to distribute the total expected responses
among the three large CMAs, the remainder of the province where these three CMAs
occurred, and all other provinces. The sample was then allocated proportionally within
the province to the remaining strata. A response rate of 80% was assumed, thus a
sample size of 50,000 would be required to obtain the 40,000 responses. With an RDD
design it is necessary to take into account that not all telephone numbers will be valid
residential numbers. An RDD sample will include a significant number of business and
non-working numbers. In addition, during the data collection process there will inevitably
be some numbers which will not be able to be resolved as being a business or a
3
Residential Telephone Services Survey, Statistics Canada, 2004.
Special Surveys Division
15
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
residential number. The sample size was increased to take all these occurrences into
account based on the experience of Statistics Canada's General Social Survey. The
resulting sample size was 120,650.
5.3
Sample selection
The sample for the provincial component of the CSGVP was generated using a refinement of
RDD sampling called the Elimination of Non-Working Banks (ENWB). Within each stratum, a list
of working banks (area code + next five digits) was compiled from telephone company
administrative files. A working bank, for the purposes of social surveys, is defined as a bank
which contains at least one working residential telephone number. Thus, all banks with only
unassigned, cell phone, non-working, or business telephone numbers are excluded from the
survey frame.
A systematic sample of banks (with replacement) was selected within each stratum. For each
selected bank, a two-digit number (00 to 99) was generated at random. This random number was
added to the bank to form a complete telephone number. This method allowed listed and unlisted
residential numbers, as well as business and non-working (i.e., not currently or never in-service)
numbers, to have a chance of being in the sample. An automated pre-dialling screening activity,
aimed at removing not-in-service and known business numbers, was performed prior to sending
the sample to the computer-assisted telephone interviewing (CATI) unit. The final sample sent to
the CATI unit consisted of 90,721 telephone numbers.
Each telephone number in the CATI sample was dialled to determine whether or not it reached a
household. If the telephone number was found to reach a household, the person answering the
telephone was asked to provide information on the individual household members. One person
in the household aged 15 or above was selected at random to complete the survey. Proxy
interviews were not accepted.
The selected respondent was asked a series of 15 questions which determined their volunteer
status. If the respondent was found to be a volunteer, they continued through the rest of the
questionnaire. On the other hand, non-volunteers were sub-sampled at a rate of 50% and only
the sub-sample continued through the remaining relevant sections of the questionnaire. At the
time the sample file was created, a flag was included which was randomly set so that it had a
50% chance of being set to one and a 50% chance of being set to zero. If a respondent was a
non-volunteer and the randomly set flag on the sample file had been set to one, then they
continued; if the flag had been set to zero, the interview ended after the series of 15 questions.
16
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
5.4
Sample size by province
The following table shows the number of telephone numbers generated for the provincial
component of the 2004 CSGVP, as well as the number of respondents before and after subsampling of non-volunteers:
Province
Newfoundland and Labrador
Prince Edward Island
Nova Scotia
New Brunswick
Quebec
Ontario
Manitoba
Saskatchewan
Alberta
British Columbia
All Provinces
Special Surveys Division
Number of
telephone numbers
generated
Number of
responses before
sub-sampling
non-volunteers
Number of
responses after
sub-sampling
non-volunteers
9,003
5,803
8,619
9,623
17,914
25,217
8,445
9,034
9,172
17,820
1,990
1,314
2,182
2,113
4,510
5,421
2,554
2,272
2,487
4,188
1,407
936
1,612
1,510
2,948
4,071
1,834
1,688
1,807
3,019
120,650
29,031
20,832
17
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
6.0
Survey methodology for the territorial (northern) component
In the three territories, the Canada Survey of Giving, Volunteering and Participating (CSGVP) was
administered between August 30, 2004 and November 15, 2004 to a sub-sample of dwellings taken from
three months of the Labour Force Survey (LFS) sample combined. The sample design of the CSGVP in
the territories is therefore closely tied to that of the LFS. The CSGVP was not collected as a true live LFS
supplement since the collection was not done at the same time as the LFS collection. As a result, the
CSGVP had to repeat the collection of the roster information as well as any LFS variables of interest.
6.1
Population coverage
The target population consisted of the population 15 years of age and older residing in Canada’s
three territories with the following exceptions:
•
•
institutionalized population
residents of Indian Reserves (with one exception, residents of the Hay River Reserve in
the Northwest Territories are included in the target population)
• full-time members of the Canadian Armed Forces
In the Yukon and Northwest Territories, only the population in selected communities is surveyed
by the LFS. For operational and cost reasons, very small communities are excluded. It is
estimated that the communities covered represent over 90% of the population aged 15 and over
in the Yukon and Northwest Territories. In Nunavut, the communities eligible for sampling cover
less than 70% of the population aged 15 and over. The estimates are, however, weighted to the
total target population aged 15 plus.
6.2
Sample design
The LFS in the north employs a multi-stage design. In the north, communities form the primary
sampling units (PSU). Sampling of PSUs is followed by sampling of households.
6.2.1
Sample rotation
The LFS design in the north employs a rotating panel design in which the sample
consists of eight panels, or rotation groups. The households in the panel are contacted
once every three months and remain in the sample for eight quarters. This results in the
household being in the sample for almost two years. The survey is conducted monthly.
One third of the quarterly sample is contacted each month, thus, 1/24th of the sample is
rotated each month.
6.2.2
Modifications to the Labour Force Survey design in
the territories for the Canada Survey of Giving,
Volunteering and Participating
The CSGVP sample included all households in the October, November and December
2004 LFS sample excluding households that were in the LFS sample for the first time.
The CSGVP used seven of the eight rotation groups in the October, November and
December 2004 LFS sample. The birth rotation group was excluded. Roster information
was collected for all members of the household and then one household member 15
years of age or older was selected at random to complete the remainder of the CSGVP
questionnaire. Proxy responses were not permitted. Unlike the provincial component, in
the territorial component there was no sub-sampling of the non-volunteers. All nonvolunteers were asked to complete all relevant sections of the questionnaire.
Special Surveys Division
19
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
6.3
Sample size
The sample consisted of the non-birth rotation groups of the October, November, December 2004
quarterly sample of the LFS. The initial sample size was 1,831. The following table gives the
breakdown by territory:
Territory
Nunavut
Northwest Territories
Yukon
Total territories
20
Initial
sample size
Number of
respondents
438
680
713
335
489
508
1,831
1,332
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
7.0
Data collection
7.1
Questionnaire design
The 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP) provides a new way
of measuring giving, volunteering and participating. It replaces the way these behaviours were
measured in the 1997 and 2000 National Survey of Giving, Volunteering and Participating.
Experiences gained from the 2000 NSGVP suggested that a number of adjustments were
required related to the questionnaire content.
In preparation for the 2004 CSGVP, consultations were held with key federal, provincial and
territorial government representatives, as well as representatives from the voluntary sector and
academics. These consultations were focused primarily on survey content and were held from
January through April 2002. Following the consultations, the steering committee members met to
discuss priorities and content issues. This meeting resulted in the development of a draft
questionnaire to be used in focus-group testing and one-on-one interviews. Qualitative testing of
content was conducted during the summer months across Canada. Changes to the survey
subsequent to the qualitative testing resulted in a pilot test in April 2003. This allowed adjustment
for any errors in the computer application, and also provided an opportunity to refine the survey
procedures.
The types of questions included in the CSGVP are divided into two major categories: those that
measure behaviours and indicate what individuals are doing in terms of their giving, volunteering
and participating, and those that measure correlates of these behaviours. This latter category
includes attitudes and motivations, as well as factors that potentially constrain or facilitate giving
and volunteering.
7.2
Supervision and quality control
All Statistics Canada interviewers are under the supervision of a staff of senior interviewers who
are responsible for ensuring that interviewers are familiar with the concepts and procedures of the
surveys to which they are assigned. Senior interviewers are also responsible for periodically
monitoring the interviewers.
Interviewers were trained on the survey content and the computer-assisted telephone
interviewing (CATI) application. In addition to classroom training, the interviewers completed a
series of mock interviews to become familiar with the survey and its concepts and definitions.
7.3
Data collection methodology
7.3.1
Provincial component
For the 10 provinces, all data were collected using computer-assisted telephone
interviewing. The CATI system has a number of generic modules which can be quickly
adapted to most types of surveys. A front-end module contains a set of standard
response codes for dealing with all possible call outcomes, as well as the associated
scripts to be read by the interviewers. The survey introduction used a standard approach
which introduces the agency, informs the respondent of the name and purpose of the
survey and the names of the survey sponsors, outlines how survey results will be used
and provides an estimated interview duration.
The random selection of one person per household was carried out at the time of the
interview. The interviewer first obtained the age, sex and relationships of everyone in the
Special Surveys Division
21
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
household. Once this information was completed, the CATI application randomly selected
one individual to be the CSGVP respondent. Respondents were informed that their
participation in the survey was voluntary, and that their information would remain strictly
confidential.
The CATI application ensured that only valid question responses were entered and that
all the correct flows were followed. Edits were built into the application to check the
consistency of responses, identify and correct outliers, and to control who gets asked
specific questions. This meant that the data was already quite “clean” at the end of the
collection process.
The cases were distributed to five Statistics Canada regional offices. The workload and
interviewing staff within each office was managed by a project manager. The automated
scheduler used by the CATI system ensured that cases were assigned randomly to
interviewers. There were a maximum of 20 call attempts per case identified as a
residential phone number; once the maximum was reached, the case was reviewed by a
senior interviewer who determined if additional calls would be made.
7.3.2
Territorial component
Collection of the CSGVP in the territories was very similar to the collection in the
provinces with the following exceptions:
•
All data were collected using a computer-assisted personal interview (CAPI)
application which allowed responses to be captured directly by the interviewer at
the time of the interview; and
•
While most interviews were collected by telephone (74%), for households without
landlines, interviews were conducted in person (26%).
7.4
Non-response
Interviewers were instructed to make all reasonable attempts to obtain a completed interview with
the randomly selected member of the household. Those who at first refused to participate were
re-contacted up to two more times to explain the importance of the survey and to encourage their
participation. For cases in which the timing of the interviewer’s call was inconvenient, an
appointment was arranged to call back at a more convenient time. For cases in which there was
no one home, numerous call backs were made.
22
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
8.0
Data processing
The main output of the Canada Survey of Giving, Volunteering and Participating (CSGVP) is a “clean”
microdata file. This chapter presents a brief summary of the processing steps involved in producing this
file.
8.1
Data capture
Responses to survey questions are captured directly by the interviewer at the time of the
interview using a computerized questionnaire. The computerized questionnaire reduces
processing time and costs associated with data entry, transcription errors and data transmission.
Some editing is done directly at the time of the interview. Where the information entered is out of
range (too large or small) of expected values, or inconsistent with the previous entries, the
interviewer is prompted, through message screens on the computer, to modify the information.
However, for some questions interviewers have the option of bypassing the edits, and of skipping
questions if the respondent does not know the answer or refuses to answer. Therefore, the
response data are subjected to further edit and imputation processes once they arrive in head
office.
8.2
Editing
The first stage of survey processing undertaken at head office was the replacement of any “outof-range” values on the data file with blanks. This process was designed to make further editing
easier.
The first type of error treated was errors in questionnaire flow, where questions which did not
apply to the respondent (and should therefore not have been answered) were found to contain
answers. In this case a computer edit automatically eliminated superfluous data by following the
flow of the questionnaire implied by answers to previous, and in some cases, subsequent
questions.
The second type of error treated involved a lack of information in questions which should have
been answered. For this type of error, a non-response or “not-stated” code was assigned to the
item.
8.3
Coding of open-ended questions
A few data items on the questionnaire were recorded by interviewers in an open-ended format,
and coded at head office. The computerized questionnaire contained a pick-list of common
organizations which was used to assist the interviewer when entering information regarding the
type of organization for which the respondent volunteered (VD_Q01) or to which the respondent
donated (GS_Q01). If the organization cited by the respondent was not on this pick-list, the
respondent was asked to provide some information regarding what the organization does. This
information was used to code the type of organization using the International Classification of
Nonprofit Organizations (ICNPO), Revision 1 (see Chapter 4.0 for further information on this
classification system).
Coding of the industry (1997 and 2002 North American Industry Classification Systems) and
occupation (1991 Standard Occupational Classification and 2001 National Occupational
Classification System – Statistics) relating to the job which the respondent had the week
preceding the interview was performed based on responses to questions LF_Q05 to LF_Q08.
Special Surveys Division
23
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
For the following six questions on the CSGVP questionnaire, the text in the “Other – specify”
write-in category was examined at head office and, where possible, coded into an existing
category:
•
FV_Q16: other volunteer activities;
•
FG_Q15: other methods in which donations were made to a charitable or non-profit
organization;
•
PA_Q11: other types of organizations to which the respondent belonged;
•
SD_Q01: religion;
•
SD_Q03: country of birth; and
•
SD_Q08: ancestral ethnicity.
8.4
Imputation
Imputation is the process that supplies valid values for those variables that have been identified
for a change either because of invalid information or because of missing information. The new
values are supplied in such a way as to preserve the underlying structure of the data and to
ensure that the resulting records will pass all required edits. In other words, the objective is not to
reproduce the true microdata values, but rather to establish internally consistent data records that
yield good aggregate estimates.
We can distinguish between three types of non-response. Complete non-response is when the
respondent does not provide the minimum set of answers. These records are dropped and
accounted for in the weighting process (see Chapter 12.0). Item non-response is when the
respondent does not provide an answer to one question, but goes on to the next question. These
are usually handled using the “not stated” code or are imputed. Finally, partial non-response is
when the respondent provides the minimum set of answers but does not finish the interview.
These records can be handled like either complete non-response or multiple item non-response.
In the case of the CSGVP, donor imputation was used to fill in missing data for some item and
partial non-response. Further information on the imputation process is given in Section 9.2.3.
8.5
Creation of derived variables
A number of data items on the microdata file have been derived by combining items on the
questionnaire in order to facilitate data analysis. Most derived variable names have a ”D” in the
fourth character position of the name. Some derived variables may have a ”G” in the fourth
character position of the name. In most cases, these are variables which have been grouped for
ease of use.
Examples of derived variables include:
•
total number of hours volunteered (VD1DHRS);
•
total number hours volunteered for the 15 organization types (VD1DTX01 to VD1DTX15 on
the master file, VD1GTX01 to VD1GTX15 on the public use microdata file (PUMF));
•
total amount of donations (GS1DATOT on the master file, GS1GATOT on the PUMF);
•
total amount of donations for the 15 organization types (GS1DAX01 to GS1DAX15 on the
master file, GS1GAX01 to GS1GAX15 on the PUMF); and
•
total amount of donations by solicitation method (FG1DA03 to FG1DA15 on the master file,
FG1GA03 to FG1GA15 on the PUMF).
Derived variables for donations were derived from the Giving (GS) file and placed on the MAIN
file (see Chapter 14.0 for further information on the file structure.)
In general, a derived variable was not calculated if any part of the equation was not answered
(i.e., don’t know, refused or not stated.) In these cases, the code assigned to the derived variable
was usually “not stated”.
24
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
8.6
Weighting
The principle behind estimation in a probability sample is that each person in the sample
“represents”, besides himself or herself, several other persons not in the sample. For example, in
a simple random 2% sample of the population, each person in the sample represents 50 persons
in the population.
The weighting phase is a step which calculates, for each record, what this number is. This weight
appears on the microdata file, and must be used to derive meaningful estimates from the survey.
For example if the number of people who had volunteered in the preceding 12 months is to be
estimated, it is done by selecting the records referring to those individuals in the sample with that
characteristic and summing the weights entered on those records.
Details of the method used to calculate these weights are presented in Chapter 12.0.
8.7
Suppression of confidential information
It should be noted that the “Public Use” Microdata Files may differ from the survey “master” files
held by Statistics Canada. These differences usually are the result of actions taken to protect the
anonymity of individual survey respondents. The most common actions are the suppression of
file variables, grouping values into wider categories, and coding specific values into the “not
stated” category. Users requiring access to information excluded from the microdata files may
purchase custom tabulations. Estimates generated will be released to the user, subject to
meeting the guidelines for analysis and release outlined in Chapter 10.0 of this document.
The survey master file includes geographic identifiers that are more explicit than the PUMF,
notably census metropolitan areas and urban centres. The PUMF does not contain any
geographic identifiers below the provincial level. The master file also includes some demographic
variables which have been excluded from the PUMF, such as ancestral ethnicity and immigration
status.
The survey master file includes certain detailed information which is included on the PUMF only
in grouped form. This includes:
•
precise age of respondent;
•
number of children aged 0 to 5 in the household (on the PUMF this has been grouped to a
yes/no variable indicating presence of children aged 0 to 5 in the household);
•
a detailed 43 category North American Industry Classification which only appears as an 18
category grouping on the PUMF;
•
country of birth, which, for the PUMF has been grouped to “Canada” and “Outside
Canada”.
As well, for certain variables that are susceptible to identifying individuals, the PUMF may have
been treated with local suppression, that is, some of the values in the master file may have been
coded as “not stated” on the PUMF.
Special Surveys Division
25
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
9.0
Data quality
9.1
Response rates
9.1.1
Response to the provincial component
The telephone resolved rate and telephone hit rate, by province, are provided in the
following table.
The telephone resolved rate is defined as the proportion of telephone numbers
confirmed, either in the pre-screening process or in the field, as being either residential or
out-of-scope (e.g., business or non-working numbers, numbers for cell phones, nonresidences or collective dwellings) as a proportion of the total number of telephone
numbers generated.
resolved rate =
number of resolved telephone numbers
number of telephone numbers generated
The hit rate is defined as the proportion of resolved telephone numbers that were
confirmed to be residential telephone numbers.
hit rate =
number of residential telephone numbers
number of resolved telephone numbers
Telephone resolved rate and hit rate by province
Telephone
numbers
generated
Telephone
numbers
sent to
collection
Telephone
numbers
resolved
in the field
Total
resolved
Newfoundland and
Labrador
9,003
5,644
4,905
8,264
Prince Edward
Island
5,803
3,697
3,242
Nova Scotia
8,619
6,049
New Brunswick
9,623
Province
Confirmed
residential
telephone
numbers
Responses
Hit
rate
(%)
91.8
2,884
1,990
34.9
5,348
92.2
2,090
1,314
39.1
5,482
8,052
93.4
3,209
2,182
39.9
6,449
5,760
8,934
92.8
3,086
2,113
34.5
Resolved
rate (%)
Quebec
17,914
14,883
13,404
16,435
91.7
8,939
4,510
54.4
Ontario
25,217
20,264
17,731
22,684
90.0
10,809
5,421
47.7
Manitoba
8,445
6,592
6,297
8,150
96.5
3,714
2,554
45.6
Saskatchewan
9,034
5,743
5,209
8,500
94.1
3,727
2,272
43.8
Alberta
9,172
7,310
6,760
8,622
94.0
4,367
2,487
50.6
17,820
14,090
13,094
16,824
94.4
8,452
4,188
50.2
120,650
90,721
81,884
111,813
92.7
51,277
29,031
45.9
British Columbia
All Provinces
Response rates are given for the provincial component of the Canada Survey of Giving,
Volunteering and Participating (CSGVP) in the following table. A respondent is defined
as a sampled person who completed the 15 questions in the Formal Volunteering (FV)
module of the questionnaire that determine whether or not the person was a volunteer.
The response rate is defined as the number of sampled persons who completed at least
Special Surveys Division
27
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
this minimum requirement divided by the number of confirmed residential telephone
numbers.
response rate =
number of respondents
number of residential telephone numbers
Response rate by province
Confirmed
residential
telephone
numbers
Province
Responses
Response
rate (%)
Newfound land and Labrador
2,884
1,990
69.0
Prince Edward Island
2,090
1,314
62.9
Nova Scotia
3,209
2,182
68.0
New Brunswick
3,086
2,113
68.5
Quebec
8,939
4,510
50.5
Ontario
10,809
5,421
50.2
Manitoba
3,714
2,554
68.8
Saskatchewan
3,727
2,272
61.0
Alberta
4,367
2,487
56.9
British Columbia
8,452
4,188
49.6
51,277
29,031
56.6
All Provinces
9.1.2
Response to the territorial component
Response rates are given for the territorial component of the CSGVP in the following
table. The same definition of respondent applies in the territories as in the provinces.
response rate =
number of respondents
number of respondents + number of non - respondents
Response rate by territory
Territory
Out of
scope
Respondents
Nonrespondents
Response
rate (%)
Yukon
713
122
508
83
86.0
Northwest Territories
680
102
489
89
84.6
Nunavut
438
67
335
36
90.3
1,831
291
1,332
208
86.5
All Territories
28
Total
sample
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
9.2
Survey errors
The estimates derived from this survey are based on a sample of households. Somewhat
different estimates might have been obtained if a complete census had been taken using the
same questionnaire, interviewers, supervisors, processing methods, etc. as those actually used in
the survey. The difference between the estimates obtained from the sample and those resulting
from a complete count taken under similar conditions, is called the sampling error of the estimate.
Errors which are not related to sampling may occur at almost every phase of a survey operation.
Interviewers may misunderstand instructions, respondents may make errors in answering
questions, the answers may be incorrectly entered on the questionnaire and errors may be
introduced in the processing and tabulation of the data. These are all examples of non-sampling
errors.
Over a large number of observations, randomly occurring errors will have little effect on estimates
derived from the survey. However, errors occurring systematically will contribute to biases in the
survey estimates. Considerable time and effort were taken to reduce non-sampling errors in the
survey. Quality assurance measures were implemented at each step of the data collection and
processing cycle to monitor the quality of the data. These measures include the use of highly
skilled interviewers, extensive training of interviewers with respect to the survey procedures and
questionnaire, observation of interviewers to detect problems of questionnaire design or
misunderstanding of instructions, procedures to ensure that data capture errors were minimized,
and coding and edit quality checks to verify the processing logic.
9.2.1
Data collection
Interviewer training consisted of a self-study of the CSGVP Interviewer’s Manual and a
review of the summary publication Caring Canadians, Involved Canadians: Highlights
from the 2000 National Survey of Giving, Volunteering and Participating, followed by two
days of classroom training. The manuals included a description of the background and
objectives of the survey, as well as a glossary of terms and a set of questions and
answers. The classroom sessions included a presentation of survey objectives, a review
of key concepts and practice time with training cases (mock interviews) using the
computer-assisted telephone interviewing (CATI) application. They also provided an
opportunity for interviewers to ask questions before the start of collection.
9.2.2
Data processing
Data processing of the CSGVP was done in a number of steps including verification,
coding, editing, imputation, estimation, and confidentiality. At each step a picture of the
output files was taken and verification was performed by comparing the files at the
current and previous step.
9.2.3
Non-response and imputation
A major source of non-sampling errors in surveys is the effect of non-response on the
survey results. The extent of non-response varies from item or partial non-response
(failure to answer just one or some questions) to total non-response. Total non-response
occurred either because the interviewer was unable to contact the respondent, because
no member of the household was able to provide the information, or because the
respondent refused to participate in the survey. Total non-response was handled by
adjusting the weight of individuals who responded to the survey to compensate for those
who did not respond.
Special Surveys Division
29
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
In most cases, item or partial non-response to the survey occurred when the respondent
did not understand or misinterpreted a question, refused to answer a question, or could
not recall the requested information. In item and partial non-response cases, for certain
variables donor imputation was performed. Most of these imputations were done in order
to provide complete data enabling the calculation of totals (e.g., total number of hours
and total value of donations). Also, the imputation helped to keep records in the sample,
even if part of the required information was not filled in by the respondent.
All imputations involved donor records that were selected using a score function. For
each item non-response or partial non-response record (also called a recipient record),
certain characteristics were compared to those from all potential donor records. When a
characteristic was the same for a donor record and the recipient record, a value was
added to the score of that donor. The donor record with the highest score was deemed
the “closest” donor and was chosen to fill in missing pieces of information of the nonrespondent. If there was more than one donor record with the highest score, a random
selection occurred. The pool of donor records was made up in such a way that the
imputed value assigned to the recipient, in conjunction with other non-imputed items from
the recipient, would still pass the edits.
Imputation was done in six steps. The first three steps related to imputation of variables
on the Main file (see Chapter 14.0 for file structure). The first step was to impute both
personal and household income. The second step was to impute the hours volunteered
by activity for the main organization. The third step was to impute the total hours
volunteered for the second and third organizations and the total hours volunteered for all
other organizations combined. The fourth step was to impute variables on the Giving
(GS) file related to amount donated. This step also included creating additional GS file
records for cases where a value for GS_Q07, Did you make any other donations in
response to this solicitation method?, was imputed as “yes”. The fifth step was to impute,
on the Main file, missing data in any of the variables indicating whether the respondent
made a donation in response to each of the 13 methods of solicitation (FG_Q03 to
FG_Q15 from the Financial Giving to Charitable Organizations (FG) section of the
questionnaire). At this stage, imputation was performed only for cases which were
already known to be givers (i.e., cases which already had a value of "yes" in at least one
of FG_Q03 to FG_Q15). This step also included creating additional GS records for cases
where one or more of FG_Q03 to FG_Q15 was imputed as "yes". The sixth step was to
impute partially completed records where the donor status could not be determined
because of missing values in FG_Q03 to FG_Q15. A total of 88 variables were imputed.
This last step again included creating additional GS file records for cases where any of
FG_Q03 to FG_Q15 was imputed as “yes”.
The following table shows the number of records imputed for some of the key variables of
the survey. The rates for the income variables are high but in 45% of the cases where
the personal income value was imputed, the respondent had reported an income range.
For the household income this percentage was 28%.
Number and percentage of records imputed for selected variables
Variable
30
Records imputed
Total records
% imputed
Personal income
8,418
22,164
38
Household income
9,131
22,164
41
Hours for organization 1
599
22,164
3
Hours for organization 2
285
22,164
1
Hours for organization 3
236
22,164
1
Hours for organizations 4+
188
22,164
1
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Donations to organizations 1 to 10
Donations to organizations 11+
16,132
93,047
17
4,867
93,047
5
The following table shows the resulting impact on the actual estimates.
Percentage of estimate originating from imputed values
Imputed
estimate
(millions)
Total
estimate
(millions)
% imputed
Hours for organization 1
70.6
1,467.8
5
Hours for organization 2
14.6
311.7
5
Hours for organization 3
7.5
117.0
6
Hours for organizations 4+
7.1
86.8
8
1,732.3
8,882.3
20
0.9
22.2
4
Variable
Amount of total donations
Number of donors
The CSGVP imputation process worked well and helped to fill incomplete responses with
the experience of other respondents with similar or identical characteristics. This adds to
the number of units used in any analysis performed by researchers.
Note that the public use microdata file does not contain any of the imputation flags. The
impact of this is an additional layer of confidentiality.
9.2.4
Measurement of sampling error
Since it is an unavoidable fact that estimates from a sample survey are subject to
sampling error, sound statistical practice calls for researchers to provide users with some
indication of the magnitude of this sampling error. This section of the documentation
outlines the measures of sampling error which Statistics Canada commonly uses and
which it urges users producing estimates from this microdata file to use also.
The basis for measuring the potential size of sampling errors is the standard error of the
estimates derived from survey results.
However, because of the large variety of estimates that can be produced from a survey,
the standard error of an estimate is usually expressed relative to the estimate to which it
pertains. This resulting measure, known as the coefficient of variation (CV) of an
estimate, is obtained by dividing the standard error of the estimate by the estimate itself
and is expressed as a percentage of the estimate.
For example, suppose that, based upon the survey results, one estimates that 54.8% of
Canadians aged 15 to 24 had done some volunteering in the preceding year, and this
estimate is found to have a standard error of 0.012. Then the coefficient of variation of
the estimate is calculated as:
⎛ 0 . 012 ⎞
⎜
⎟ X 100 % = 2 . 2 %
⎝ 0 . 548 ⎠
There is more information on the calculation of coefficients of variation in Chapter 11.0.
Special Surveys Division
31
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
10.0 Guidelines for tabulation, analysis and release
This chapter of the documentation outlines the guidelines to be adhered to by users tabulating, analyzing,
publishing or otherwise releasing any data derived from the survey microdata files. With the aid of these
guidelines, users of microdata should be able to produce the same figures as those produced by
Statistics Canada and, at the same time, will be able to develop currently unpublished figures in a manner
consistent with these established guidelines.
10.1 Rounding guidelines
In order that estimates for publication or other release derived from these microdata files
correspond to those produced by Statistics Canada, users are urged to adhere to the following
guidelines regarding the rounding of such estimates:
a) Estimates in the main body of a statistical table are to be rounded to the nearest hundred
units using the normal rounding technique. In normal rounding, if the first or only digit to
be dropped is 0 to 4, the last digit to be retained is not changed. If the first or only digit to
be dropped is 5 to 9, the last digit to be retained is raised by one. For example, in normal
rounding to the nearest 100, if the last two digits are between 00 and 49, they are
changed to 00 and the preceding digit (the hundreds digit) is left unchanged. If the last
digits are between 50 and 99 they are changed to 00 and the preceding digit is
incremented by 1.
b) Marginal sub-totals and totals in statistical tables are to be derived from their
corresponding unrounded components and then are to be rounded themselves to the
nearest 100 units using normal rounding.
c) Averages, proportions, rates and percentages are to be computed from unrounded
components (i.e. numerators and/or denominators) and then are to be rounded
themselves to one decimal using normal rounding. In normal rounding to a single digit, if
the final or only digit to be dropped is 0 to 4, the last digit to be retained is not changed. If
the first or only digit to be dropped is 5 to 9, the last digit to be retained is increased by 1.
d) Sums and differences of aggregates (or ratio) are to be derived from their corresponding
unrounded components and then are to be rounded themselves to the nearest 100 units
(or the nearest one decimal) using normal rounding.
e) In instances where, due to technical or other limitations, a rounding technique other than
normal rounding is used resulting in estimates to be published or otherwise released
which differ from corresponding estimates published by Statistics Canada, users are
urged to note the reason for such differences in the publication or release document(s).
f)
Under no circumstances are unrounded estimates to be published or otherwise released
by users. Unrounded estimates imply greater precision than actually exists.
10.2 Sample weighting guidelines for tabulation
The sample design used for the Canada Survey of Giving, Volunteering and Participating
(CSGVP) was not self-weighting. When producing simple estimates, including the production of
ordinary statistical tables, users must apply the proper survey weight.
If proper weights are not used, the estimates derived from the microdata files cannot be
considered to be representative of the survey population, and will not correspond to those
produced by Statistics Canada.
Special Surveys Division
33
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Users should also note that some software packages may not allow the generation of estimates
that exactly match those available from Statistics Canada, because of their treatment of the
weight field.
10.3 Definitions of types of estimates: categorical and
quantitative
Before discussing how the CSGVP data can be tabulated and analyzed, it is useful to describe
the two main types of point estimates of population characteristics which can be generated from
the microdata file for the CSGVP.
10.3.1
Categorical estimates
Categorical estimates are estimates of the number, or percentage of the surveyed
population possessing certain characteristics or falling into some defined category. The
number of Canadians who volunteered, or the number of Canadians who made financial
donations are examples of such estimates. An estimate of the number of persons
possessing a certain characteristic may also be referred to as an estimate of an
aggregate.
Examples of categorical questions:
Q:
R:
Q:
R:
10.3.2
In the past 12 months, did you do any of the following activities without
pay on behalf of a group or an organization? This includes any unpaid
help you provided to schools, religious organizations, sports or
community associations. Did you do any: … teaching, educating or
mentoring?
Yes / No
In the past 12 months, did you make a charitable donation: … by
responding to a request through the mail?
Yes / No
Quantitative estimates
Quantitative estimates are estimates of totals or of means, medians and other measures
of central tendency of quantities based upon some or all of the members of the surveyed
population. They also specifically involve estimates of the form
Xˆ / Yˆ where X̂ is an
estimate of surveyed population quantity total and Yˆ is an estimate of the number of
persons in the surveyed population contributing to that total quantity.
An example of a quantitative estimate is the average number of hours contributed by
volunteers. The numerator is an estimate of the total number of hours volunteered and
its denominator is the number of persons who volunteered.
Examples of quantitative questions:
Q:
34
R:
In the past 12 months, how many hours did you spend on unpaid
activities for this organization?
|_|_|_|_| hours
Q:
R:
What was the amount of the donation to this organization?
|_|_|_|_|_| dollars
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
10.3.3
Tabulation of categorical estimates
Estimates of the number of people with a certain characteristic can be obtained from the
microdata file by summing the final weights of all records possessing the characteristic(s)
of interest. Proportions and ratios of the form
Xˆ / Yˆ are obtained by
a) summing the final weights of records having the characteristic of interest for the
numerator ( X̂ ),
b) summing the final weights of records having the characteristic of interest for the
denominator ( Yˆ ), then
c) dividing estimate a) by estimate b) ( Xˆ
10.3.4
/ Yˆ ).
Tabulation of quantitative estimates
Estimates of quantities can be obtained from the microdata file by multiplying the value of
the variable of interest by the final weight for each record, then summing this quantity
over all records of interest. For example, to obtain an estimate of the total number of
hours volunteered by persons aged 65 and over, multiply the value reported in
VD1DHRS (hours volunteered) by the final weight for the record, then sum this value
over all records with DH1GAGE = 6 (age group 65 and over).
Xˆ / Yˆ , the numerator ( X̂ ) is calculated as for
a quantitative estimate and the denominator ( Yˆ ) is calculated as for a categorical
To obtain a weighted average of the form
estimate. For example, to estimate the average number of hours volunteered by those
aged 65 and over,
a) estimate the total number of hours volunteered ( X̂ ) as described above,
b) estimate the number of people in this category ( Yˆ ) in this category by summing
the final weights of all records with DH1GAGE = 6, then
c) divide estimate a) by estimate b) ( Xˆ
/ Yˆ ).
10.4 Guidelines for statistical analysis
The CSGVP is based upon a complex sample design, with stratification, multiple stages of
selection, and unequal probabilities of selection of respondents. Using data from such complex
surveys presents problems to analysts because the survey design and the selection probabilities
affect the estimation and variance calculation procedures that should be used. In order for survey
estimates and analyses to be free from bias, the survey weights must be used.
While many analysis procedures found in statistical packages allow weights to be used, the
meaning or definition of the weight in these procedures may differ from that which is appropriate
in a sample survey framework, with the result that while in many cases the estimates produced by
the packages are correct, the variances that are calculated are poor. Approximate variances for
simple estimates such as totals, proportions and ratios (for qualitative variables) can be derived
using the accompanying Approximate Sampling Variability Tables.
For other analysis techniques (for example linear regression, logistic regression and analysis of
variance), a method exists which can make the variances calculated by the standard packages
more meaningful, by incorporating the unequal probabilities of selection. The method rescales
the weights so that there is an average weight of one.
Special Surveys Division
35
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
For example, suppose that analysis of all male respondents is required. The steps to rescale the
weights are as follows:
1) select all respondents from the file who reported RESPSEX = male;
2) calculate the AVERAGE weight for these records by summing the original person weights
from the microdata file for these records and then dividing by the number of respondents
who reported RESPSEX = male;
3) for each of these respondents, calculate a RESCALED weight equal to the original
person weight divided by the AVERAGE weight;
4) perform the analysis for these respondents using the RESCALED weight.
However, because the stratification and clustering of the sample’s design are still not taken into
account, the variance estimates calculated in this way are likely to be under-estimates.
The calculation of more precise variance estimates requires detailed knowledge of the design of
the survey. Such detail cannot be given in this microdata file because of confidentiality.
Variances that take the complete sample design into account can be calculated for many
statistics by Statistics Canada on a cost-recovery basis.
10.5 Coefficient of variation release guidelines
Before releasing and/or publishing any estimates from the CSGVP, users should first determine
the quality level of the estimate. The quality levels are acceptable, marginal and unacceptable.
Data quality is affected by both sampling and non-sampling errors as discussed in Chapter 9.0.
However for this purpose, the quality level of an estimate will be determined only on the basis of
sampling error as reflected by the coefficient of variation as shown in the table below.
Nonetheless users should be sure to read Chapter 9.0 to be more fully aware of the quality
characteristics of these data.
First, the number of respondents who contribute to the calculation of the estimate should be
determined. If this number is less than 30, the weighted estimate should be considered to be of
unacceptable quality.
For weighted estimates based on sample sizes of 30 or more, users should determine the
coefficient of variation of the estimate and follow the guidelines below. These quality level
guidelines should be applied to rounded weighted estimates.
All estimates can be considered releasable. However, those of marginal or unacceptable quality
level must be accompanied by a warning to caution subsequent users.
36
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Quality Level Guidelines
Quality level of estimate
Guidelines
1) Acceptable
Estimates have
a sample size of 30 or more, and
low coefficients of variation in the range of 0.0% to 16.5%.
No warning is required.
2) Marginal
Estimates have
a sample size of 30 or more, and
high coefficients of variation in the range of 16.6% to 33.3%.
Estimates should be flagged with the letter M (or some similar
identifier). They should be accompanied by a warning to caution
subsequent users about the high levels of error associated with the
estimates.
3) Unacceptable
Estimates have
a sample size of less than 30, or
very high coefficients of variation in excess of 33.3%.
Statistics Canada recommends not to release estimates of
unacceptable quality. However, if the user chooses to do so then
estimates should be flagged with the letter U (or some similar
identifier) and the following warning should accompany the
estimates:
“Please be warned that these estimates [flagged with the letter U]
do not meet Statistics Canada’s quality standards. Conclusions
based on these data will be unreliable, and most likely invalid.”
Special Surveys Division
37
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
10.6 Release cut-offs for the Canada Survey of Giving,
Volunteering and Participating
The following table provides an indication of the precision of population estimates as it shows the
release cut-offs associated with each of the three quality levels presented in the previous section.
These cut-offs are derived from the coefficient of variation (CV) tables discussed in Chapter 11.0.
For example, the table shows that the quality of a weighted estimate of 10,000 people possessing
a given characteristic in Newfoundland and Labrador is marginal.
Note that these cut-offs apply to estimates of population totals only. To estimate ratios, users
should not use the numerator value (nor the denominator) in order to find the corresponding
quality level. Rule 4 in Section 11.1 and Example 4 in Section 11.1.1 explains the correct
procedure to be used for ratios.
Province / Territories
Acceptable CV
0.0% to 16.5%
Newfoundland and Labrador
15,000
& over
4,000
to <
15,000
under
4,000
6,000
& over
1,500
to <
6,000
under
1,500
22,500
& over
5,500
to <
22,500
under
5,500
Prince Edward Island
Nova Scotia
New Brunswick
Unacceptable CV
> 33.3%
19,500
& over
5,000
to <
19,500
under
5,000
Quebec
113,000
& over
28,000
to <
113,000
under
28,000
Ontario
130,000
& over
32,000
to <
130,000
under
32,000
Manitoba
24,500
& over
6,000
to <
24,500
under
6,000
Saskatchewan
25,000
& over
6,000
to <
25,000
under
6,000
Alberta
73,000
& over
18,500
to <
73,000
under
18,500
British Columbia
60,000
& over
15,000
to <
60,000
under
15,000
Provinces
96,500
& over
23,500
to <
96,500
under
23,500
Territories
3,500
& over
1,000
to <
3,500
under
1,000
96,000
& over
23,500
to <
96,000
under
23,500
Canada
38
Marginal CV
16.6% to 33.3%
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
11.0 Approximate sampling variability tables
In order to supply coefficients of variation (CVs) which would be applicable to a wide variety of categorical
estimates produced from this microdata file and which could be readily accessed by the user, a set of
Approximate Sampling Variability Tables has been produced. These CV tables allow the user to obtain
an approximate coefficient of variation based on the size of the estimate calculated from the survey data.
The coefficients of variation are derived using the variance formula for simple random sampling and
incorporating a factor which reflects the multi-stage, clustered nature of the sample design. This factor,
known as the design effect, was determined by first calculating design effects for a wide range of
characteristics and then choosing from among these a conservative value (usually the 75th percentile) to
be used in the CV tables which would then apply to the entire set of characteristics.
The table below shows the conservative value of the design effects as well as sample sizes and
population counts by province, which were used to produce the Approximate Sampling Variability Tables
for the 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP).
Province/Territories
Design effect
Sample size
Population
Newfoundland and Labrador
1.37
1,407
440,863
Prince Edward Island
1.39
936
115,184
Nova Scotia
1.31
1,612
779,570
New Brunswick
1.33
1,510
622,946
Quebec
1.49
2,948
6,211,020
Ontario
1.45
4,071
10,068,734
Manitoba
1.37
1,834
921,621
Saskatchewan
1.49
1,688
789,055
Alberta
1.44
1,807
2,573,431
British Columbia
1.44
3,019
3,498,788
Provinces
2.11
20,832
26,021,212
Territories
1.83
1,332
71,962
Canada
2.23
22,164
26,093,174
All coefficients of variation in the Approximate Sampling Variability Tables are approximate and,
therefore, unofficial. Estimates of actual variance for specific variables may be obtained from Statistics
Canada on a cost-recovery basis. Since the approximate CV is conservative, the use of actual variance
estimates may cause the estimate to be switched from one quality level to another. For instance a
marginal estimate could become acceptable based on the exact CV calculation.
Remember:
If the number of observations on which an estimate is based is less than 30, the weighted
estimate is most likely unacceptable and Statistics Canada recommends not releasing such
an estimate, regardless of the value of the coefficient of variation.
Special Surveys Division
39
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
11.1 How to use the coefficient of variation tables for
categorical estimates
The following rules should enable the user to determine the approximate coefficients of variation
(CVs) from the Approximate Sampling Variability Tables for estimates of the number, proportion
or percentage of the surveyed population possessing a certain characteristic and for ratios and
differences between such estimates.
Rule 1:
Estimates of numbers of persons possessing a characteristic (aggregates)
The coefficient of variation depends only on the size of the estimate itself. On the Approximate
Sampling Variability Table for the appropriate geographic area, locate the estimated number in
the left-most column of the table (headed “Numerator of Percentage”) and follow the asterisks (if
any) across to the first figure encountered. This figure is the approximate coefficient of variation.
Rule 2:
Estimates of proportions or percentages of persons possessing a characteristic
The coefficient of variation of an estimated proportion or percentage depends on both the size of
the proportion or percentage and the size of the total upon which the proportion or percentage is
based. Estimated proportions or percentages are relatively more reliable than the corresponding
estimates of the numerator of the proportion or percentage, when the proportion or percentage is
based upon a sub-group of the population. For example, the proportion of volunteers who
provided health care or support including companionship is more reliable than the estimated
number of volunteers who provided health care or support including companionship. (Note that in
the tables the coefficients of variation decline in value reading from left to right).
When the proportion or percentage is based upon the total population of the geographic area
covered by the table, the CV of the proportion or percentage is the same as the CV of the
numerator of the proportion or percentage. In this case, Rule 1 can be used.
When the proportion or percentage is based upon a subset of the total population (e.g., those in a
particular sex or age group), reference should be made to the proportion or percentage (across
the top of the table) and to the numerator of the proportion or percentage (down the left side of
the table). The intersection of the appropriate row and column gives the coefficient of variation.
Rule 3:
Estimates of differences between aggregates or percentages
The standard error of a difference between two estimates is approximately equal to the square
root of the sum of squares of each standard error considered separately. That is, the standard
(
)
error of a difference dˆ = Xˆ 1 − Xˆ 2 is
σ
dˆ
(Xˆ 1α 1 )2 + (Xˆ 2α 2 )2
where X̂ 1 is estimate 1, X̂ 2 is estimate 2, and α
1
and α
2
are the coefficients of variation of
X̂ 1 and X̂ 2 respectively. The coefficient of variation of dˆ is given by σdˆ / dˆ . This formula is
accurate for the difference between separate and uncorrelated characteristics, but is only
approximate otherwise.
40
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Rule 4:
Estimates of ratios
In the case where the numerator is a subset of the denominator, the ratio should be converted to
a percentage and Rule 2 applied. This would apply, for example, to the case where the
denominator is the number of persons with a university degree and the numerator is the number
of volunteers with a university degree.
In the case where the numerator is not a subset of the denominator, as for example, the ratio of
the number of volunteers with a university degree as compared to the number of volunteers
without a university degree, the standard error of the ratio of the estimates is approximately equal
to the square root of the sum of squares of each coefficient of variation considered separately
(
)
multiplied by R̂ . That is, the standard error of a ratio Rˆ = Xˆ 1 / Xˆ 2 is
σ Rˆ = Rˆ α12 + α 2 2
where
α1
and
α2
are the coefficients of variation of X̂ 1 and X̂ 2 respectively. The coefficient of
variation of R̂ is given by
σ Rˆ / Rˆ .
The formula will tend to overstate the error if X̂ 1 and X̂ 2 are
positively correlated and understate the error if X̂ 1 and X̂ 2 are negatively correlated.
Rule 5:
Estimates of differences of ratios
In this case, Rules 3 and 4 are combined. The CVs for the two ratios are first determined using
Rule 4, and then the CV of their difference is found using Rule 3.
11.1.1
Examples of using the coefficient of variation
tables for categorical estimates
The following examples, based on the 2004 CSGVP, are included to assist users in
applying the foregoing rules.
Example 1:
Estimates of numbers of persons possessing a characteristic a
(aggregates)
Suppose that a user estimates that 5,615,215 men were volunteers during the reference
period. How does the user determine the coefficient of variation of this estimate?
1) Refer to the coefficient of variation table for CANADA.
2) The estimated aggregate (5,615,215) does not appear in the left-hand column (the
“Numerator of Percentage” column), so it is necessary to use the figure closest to it,
namely 6,000,000.
3) The coefficient of variation for an estimated aggregate is found by referring to the first
non-asterisk entry on that row, namely, 1.8%.
4) So the approximate coefficient of variation of the estimate is 1.8%. The finding that
there were 5,615,215 (to be rounded according to the rounding guidelines in Section
10.1) male volunteers during the reference period is publishable with no
qualifications.
Special Surveys Division
41
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Example 2:
Estimates of proportions or percentages of persons possessing a
characteristic
Suppose that the user estimates that 1,605,006 / 5,615,215 = 28.6% of men who
volunteered did some teaching, educating or mentoring. How does the user determine
the coefficient of variation of this estimate?
1) Refer to the coefficient of variation table for CANADA.
2) Because the estimate is a percentage which is based on a subset of the total
population (i.e., men who were volunteers), it is necessary to use both the
percentage (28.6%) and the numerator portion of the percentage (1,605,006) in
determining the coefficient of variation.
3) The numerator, 1,605,006 does not appear in the left-hand column (the “Numerator
of Percentage” column) so it is necessary to use the figure closest to it, namely
1,500,000. Similarly, the percentage estimate does not appear as any of the column
headings, so it is necessary to use the percentage closest to it, 30.0%.
4) The figure at the intersection of the row and column used, namely 3.5% is the
coefficient of variation to be used.
5) So the approximate coefficient of variation of the estimate is 3.5%. The finding that
28.6% of men who volunteered did some teaching, educating or mentoring can be
published with no qualifications.
Example 3:
Estimates of differences between aggregates or percentages
Suppose that a user estimates that 1,979,228 / 6,193,361 = 32.0% of women who
volunteered did some teaching, educating or mentoring, while
1,605,006 / 5,615,215 = 28.6% of men who volunteered did some teaching, educating or
mentoring. How does the user determine the coefficient of variation of the difference
between these two estimates?
1) Using the CANADA coefficient of variation table in the same manner as described in
Example 2 gives the CV of the estimate for women as 3.0%, and the CV of the
estimate for men as 3.5%.
(
)
2) Using Rule 3, the standard error of a difference dˆ = Xˆ 1 − Xˆ 2 is
σ dˆ =
(Xˆ α ) + (Xˆ α )
2
1
1
2
2
2
where X̂ 1 is estimate 1 (women), X̂ 2 is estimate 2 (men), and
α1
and
α2
are the
coefficients of variation of X̂ 1 and X̂ 2 respectively.
That is, the standard error of the difference
σ dˆ =
=
d̂ = 0.320 – 0.286 = 0.034 is
[(0.320 )(0.030 )]2 + [(0.286 )(0.035 )]2
(0.0000921 ) + (0.0001002 )
= 0 .014
42
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
3) The coefficient of variation of
d̂ is given by σ dˆ / dˆ = 0.014 / 0.034 = 0.412
4) So the approximate coefficient of variation of the difference between the estimates is
41.2%. The difference between the estimates is considered unacceptable and
Statistics Canada recommends this estimate not be released. However, should the
user choose to do so, the estimate should be flagged with the letter U (or some
similar identifier) and be accompanied by a warning to caution subsequent users
about the high levels of error associated with the estimate.
Example 4:
Estimates of ratios
Suppose that the user estimates that 1,979,228 women who volunteered did some
teaching, educating or mentoring on behalf of an organization, while 1,605,006 men who
volunteered did some teaching, educating or mentoring. The user is interested in
comparing the estimate of women versus that of men in the form of a ratio. How does
the user determine the coefficient of variation of this estimate?
1) First of all, this estimate is a ratio estimate, where the numerator of the estimate ( X̂ 1 )
is the number of female volunteers who did some teaching, educating or mentoring
on behalf of an organization. The denominator of the estimate ( X̂ 2 ) is the number of
male volunteers who did some teaching, educating or mentoring on behalf of an
organization.
2) Refer to the coefficient of variation table for CANADA.
3) The numerator of this ratio estimate is 1,979,228. The figure closest to it is
2,000,000. The coefficient of variation for this estimate is found by referring to the
first non-asterisk entry on that row, namely, 3.4%.
4) The denominator of this ratio estimate is 1,605,006. The figure closest to it is
1,500,000. The coefficient of variation for this estimate is found by referring to the
first non-asterisk entry on that row, namely, 4.0%
5) So the approximate coefficient of variation of the ratio estimate is given by Rule 4,
which is
α Rˆ = α 1 2 + α 2 2
where α 1 and
That is,
α2
are the coefficients of variation of X̂ 1 and X̂ 2 respectively.
α Rˆ =
(0.034 )2 + (0.040 )2
= 0.001156 + 0.0016
= 0.052
6) The obtained ratio of female versus male volunteers who did some teaching,
educating or mentoring on behalf of an organization is 1,979,228 / 1,605,006 which is
1.23 (to be rounded according to the rounding guidelines in Section 10.1). The
coefficient of variation of this estimate is 5.2%, which makes the estimate releasable
with no qualifications.
Special Surveys Division
43
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Example 5:
Estimates of differences of ratios
Suppose that the user estimates that the ratio of female volunteers to male volunteers is
1.039 for ages 15 to 24 while it is 1.169 for ages 55 and over. The user is interested in
comparing the two ratios to see if there is a statistical difference between them. How
does the user determine the coefficient of variation of the difference?
1) First calculate the approximate coefficient of variation for the 15 to 24 age group ratio
( R̂1 ) and the 55 and over age group ratio ( R̂2 ) as in Example 4. The approximate
CV for the 15 to 24 age group ratio is 7.07% and 5.66% for ages 55 and over.
2) Using Rule 3, the standard error of a difference ( dˆ = Rˆ1 − Rˆ 2 ) is
σ dˆ =
where
α1
and
α2
(Rˆ α ) + (Rˆ α )
2
1
1
2
2
2
are the coefficients of variation of R̂1 and R̂2 respectively. That
is, the standard error of the difference
σ dˆ =
=
d̂ = 1.039 – 1.169 = -0.13 is
[(1.039 )(0.0707 )]2 + [(1.169 )(0.0566 )]2
(0.005396 ) + (0.004378 )
= 0 .099
3) The coefficient of variation of
d̂ is given by σ dˆ / dˆ = 0.099 / (-0.13) = -0.762.
4) So the approximate coefficient of variation of the difference between the estimates is
76.2%. The estimate of the difference between the estimates is considered
unacceptable and Statistics Canada recommends this estimate not be released.
However, should the user choose to do so, the estimate should be flagged with the
letter U (or some similar identifier) and be accompanied by a warning to caution
subsequent users about the high levels of error, associated with the estimate.
11.2 How to use the coefficient of variation tables to obtain
confidence limits
Although coefficients of variation are widely used, a more intuitively meaningful measure of
sampling error is the confidence interval of an estimate. A confidence interval constitutes a
statement on the level of confidence that the true value for the population lies within a specified
range of values. For example a 95% confidence interval can be described as follows:
If sampling of the population is repeated indefinitely, each sample leading to a new
confidence interval for an estimate, then in 95% of the samples the interval will cover the
true population value.
Using the standard error of an estimate, confidence intervals for estimates may be
obtained under the assumption that under repeated sampling of the population, the
various estimates obtained for a population characteristic are normally distributed about
the true population value. Under this assumption, the chances are about 68 out of 100
that the difference between a sample estimate and the true population value would be
less than one standard error, about 95 out of 100 that the difference would be less than
44
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
two standard errors, and about 99 out of 100 that the difference would be less than three
standard errors. These different degrees of confidence are referred to as the confidence
levels.
Confidence intervals for an estimate, X̂ , are generally expressed as two numbers, one
(
)
below the estimate and one above the estimate, as Xˆ − k , Xˆ + k where k is
determined depending upon the level of confidence desired and the sampling error of the
estimate.
Confidence intervals for an estimate can be calculated directly from the Approximate
Sampling Variability Tables by first determining from the appropriate table the coefficient
of variation of the estimate X̂ , and then using the following formula to convert to a
confidence interval ( CI xˆ ):
(
CI xˆ = Xˆ − tXˆ α xˆ , Xˆ + tXˆ α xˆ
)
where α x̂ is the determined coefficient of variation of X̂ , and
t
t
t
t
Note:
= 1 if a 68% confidence interval is desired;
= 1.6 if a 90% confidence interval is desired;
= 2 if a 95% confidence interval is desired;
= 2.6 if a 99% confidence interval is desired.
Release guidelines which apply to the estimate also apply to the confidence
interval. For example, if the estimate is not releasable, then the confidence
interval is not releasable either.
11.2.1
Example of using the coefficient of variation tables
to obtain confidence limits
A 95% confidence interval for the estimated proportion of male volunteers who did some
teaching, educating or mentoring (from Example 2, Section 11.1.1) would be calculated
as follows:
X̂ =
28.6% (or expressed as a proportion 0.286)
t
=
2
α x̂
=
3.5% (0.035 expressed as a proportion) is the coefficient of variation of
this estimate as determined from the tables.
CI xˆ = {0.286 – (2) (0.286) (0.035), 0.286 + (2) (0.286) (0.035)}
CI xˆ = {0.286 – 0.020, 0.286 + 0.020}
CI xˆ = {0.266, 0.306}
With 95% confidence it can be said that between 26.6% and 30.6% of male volunteers
did some teaching, educating or mentoring.
Special Surveys Division
45
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
11.3 How to use the coefficient of variation tables to do a t-test
Standard errors may also be used to perform hypothesis testing, a procedure for distinguishing
between population parameters using sample estimates. The sample estimates can be numbers,
averages, percentages, ratios, etc. Tests may be performed at various levels of significance,
where a level of significance is the probability of concluding that the characteristics are different
when, in fact, they are identical.
Let X̂ 1 and X̂ 2 be sample estimates for two characteristics of interest. Let the standard error on
the difference X̂ 1 − X̂ 2 be
If t =
Xˆ 1 − Xˆ 2
σ dˆ
σ d̂ .
is between -2 and 2, then no conclusion about the difference between the
characteristics is justified at the 5% level of significance. If however, this ratio is smaller than -2
or larger than +2, the observed difference is significant at the 0.05 level. That is to say that the
difference between the estimates is significant.
11.3.1
Example of using the coefficient of variation tables
to do a t-test
Let us suppose that the user wishes to test, at 5% level of significance, the hypothesis
that there is no difference between the proportion of female volunteers who did some
teaching, educating or mentoring and the proportion male volunteers who did some
teaching, educating or mentoring. From Example 3, Section 11.1.1, the standard error of
the difference between these two estimates was found to be 0.014. Hence,
t=
X̂ 1 − X̂ 2
σ d̂
=
0.320 − 0.286 0.034
=
= 2.43
0.014
0.014
Since t = 2.43 is greater than 2, it must be concluded that there is a significant difference
between the two estimates at the 0.05 level of significance.
11.4 Coefficients of variation for quantitative estimates
For quantitative estimates, special tables would have to be produced to determine their sampling
error. Since most of the variables for the CSGVP are primarily categorical in nature, this has not
been done.
As a general rule, however, the coefficient of variation of a quantitative total will be larger than the
coefficient of variation of the corresponding category estimate (i.e., the estimate of the number of
persons contributing to the quantitative estimate). If the corresponding category estimate is not
releasable, the quantitative estimate will not be either. For example, the coefficient of variation of
the number of hours volunteered for arts and culture organizations would be greater than the
coefficient of variation of the corresponding proportion of volunteers who volunteered for arts and
culture organizations. Hence, if the coefficient of variation of the proportion is unacceptable
(making the proportion not releasable), then the coefficient of variation of the corresponding
quantitative estimate will also be unacceptable (making the quantitative estimate not releasable).
Coefficients of variation of such estimates can be derived as required for a specific estimate using
a technique known as pseudo replication. This involves dividing the records on the microdata
46
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
files into subgroups (or replicates) and determining the variation in the estimate from replicate to
replicate. Users wishing to derive coefficients of variation for quantitative estimates may contact
Statistics Canada for advice on the allocation of records to appropriate replicates and the
formulae to be used in these calculations.
11.5 Coefficient of variation tables
Refer to CSGVP2004_CVTabsE.pdf for the coefficient of variation tables.
Special Surveys Division
47
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
12.0 Weighting
A statistical weight was placed on each record of the data file. This weight indicates the number of
persons in the population represented by the sampled unit.
Since the 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP) was conducted as a
Random Digit Dialling (RDD) survey in the 10 provinces while it used a sub-sample of the Labour Force
Survey (LFS) sample in the three territories, two different sets of weighting procedures were used.
12.1 Weighting for the provincial component
The weighting for the provincial component consisted of several steps:
•
•
•
•
•
•
•
•
•
calculation of the basic telephone weight;
an adjustment for unresolved telephone numbers;
dropping out-of-scope records;
an adjustment for the number of telephone lines in the household;
adjustments for non-response (household level and person level);
an adjustment for selecting only one person from the household;
an adjustment for sub-sampling non-volunteers;
an adjustment for outliers; and
an adjustment to make the population estimates consistent with known province-age-sex
totals from the Census projected population counts for persons 15 years of age and over.
The details of these steps follow.
1. Calculation of the basic telephone weight
The initial weight is the inverse of the probability of selection of the telephone number, calculated
as follows within each stratum:
⎛ total number of possible telephone numbers from working banks ⎞
⎟⎟
w1 = ⎜⎜
number
of
sampled
telephone
numbers
⎝
⎠
There were 120,650 phone numbers selected in the sample.
2. Adjustment for unresolved telephone numbers
Before data collection, the 120,650 phone numbers underwent a screening process; 9,154
business numbers and 20,775 non-working numbers were dropped, leaving 90,721 telephone
numbers for data collection. Each of the remaining records either had an initial status equal to
residential or the initial status was unknown.
At the end of the data collection period, call history information obtained during collection was
used to determine the final status of each record. Each unit was identified as out-of-scope, inscope or unresolved. The weights of the resolved and out-of-scope records were adjusted to
account for the unresolved records and the unresolved records were dropped. The adjustment
was performed at the stratum level separately for those with initial status of residential and those
with initial status unknown (see Section 5.2 for description of strata). A total of 8,837 unresolved
records were dropped, leaving 81,884 records.
The weights were adjusted as follows within each stratum and initial status:
Special Surveys Division
49
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
⎛ ∑ w1 for resolved telephone numbers + ∑ w1 for unresolved telephone numbers ⎞
⎟
w 2 = w1 * ⎜
⎜
⎟
w
for
resolved
telephone
numbers
∑
1
⎝
⎠
3. Dropping out-of-scope telephone numbers
Phone numbers that were resolved after collection to be non-working or otherwise out-of-scope
(businesses, cell phones, non-residences, collective dwellings, etc.) were dropped. A total of
51,277 records remained at this point.
4. Adjustment for missing number of telephone lines (first household level non-response
adjustment)
In order to convert the telephone level weight calculated in Step 2 into a household level weight, it
was necessary to divide the telephone weight by the number of telephone lines associated with
the household. There are cases where the number of lines cannot be derived because of either
item non-response or total household non-response. In the case of item non-response, the
number of lines was imputed to one. The remaining cases where the number of telephone lines
could not be derived were dropped and the weights of the retained units were inflated to
compensate for the dropped records.
As a result of a non-response study, it was discovered that those cases who eventually
responded, but had at least one refusal or in-progress language barrier code in the history of
calls, had much lower volunteer rates than other cases. Adjustment groups were formed by
splitting each stratum into groups based on the presence of a refusal and/or language barrier.
The weights were adjusted as follows within each stratum and refusal / language barrier group:
⎛ ∑ w2 for households with number of lines + ∑ w 2 for households missing number of lines ⎞
⎟
w3 = w2 * ⎜
⎜
⎟
w
for
households
with
number
of
lines
∑ 2
⎝
⎠
A total of 33,714 records remained.
5. Adjustment for number of telephone lines in the household
Weights for households with more than one telephone line (with different telephone numbers)
were adjusted downwards to account for the fact that such households have a higher probability
of being selected. The telephone weight was divided by the number of lines in the household.
The maximum adjustment was capped at four to prevent outliers. At this stage the telephone
weight becomes the household weight.
The weights were adjusted as follows:
⎛
⎞
w3
⎟
w4 = ⎜
⎜ number of in - scope telephone lines in the household ⎟
⎝
⎠
6. Adjustment for household non-response (second household level non-response
adjustment)
This step accounts for the remaining non-responding households, i.e., those for whom the
number of telephone lines in the household could be derived. The weights were inflated, within
stratum, to compensate for non-responding households. Non-responding households were
dropped at this step, leaving 32,464 records.
50
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
The weights were adjusted as follows within each stratum:
⎛ ∑ w4 for household respondent s + ∑ w4 for household non - respondent s ⎞
⎟
w5 = w4 * ⎜
⎜
⎟
w
for
household
respondent
s
∑
4
⎝
⎠
7. Adjustment for sampling only one person in the household (aged 15 or over)
The household weight calculated in Step 6 was multiplied by the number of members in the
household aged 15 or over. This adjustment was capped at five to prevent outliers. After this
step, the weight changes from representing households to representing persons.
The weights were adjusted as follows:
w6 = w5 * (number of household members aged 15 + )
8. Adjustment for person level non-response
The weights were then inflated to compensate for non-responding persons. This adjustment was
done within stratum, age group and sex, and non-responding persons were dropped, leaving
29,031 records.
The weights were adjusted as follows within each stratum, age group and sex:
⎛ ∑ w6 for person respondents + ∑ w6 for person non - respondents ⎞
⎟
w7 = w6 * ⎜
⎜
⎟
w
for
person
respondent
s
∑
6
⎝
⎠
9. Adjustment for sub-sampling non-volunteers
The weighted sub-sampling rate for non-volunteers was calculated within each stratum, as
follows, using the weighted counts from the previous step:
⎛
⎞
∑ w7 for selected non − volunteers
⎟
Weighted sub − sampling rate = ⎜
⎜ ∑ w7 for selected non - volunteers + ∑ w7 for non - selected non − volunteers ⎟
⎝
⎠
The inverse of this rate was multiplied by the weights for the selected non-volunteers and the
non-selected non-volunteers were dropped. In effect, the weights of the selected non-volunteers
were approximately doubled to account for the non-volunteers who were not selected. The
theoretical rate of sub-sampling non-volunteers was 50%, but the actual weighted sub-sampling
rate within each stratum ranged from 46.5% to 59.0%.
For non-volunteers, the weights were adjusted as follows within each stratum:
⎛
⎞
w7
⎟⎟
w8 = ⎜⎜
⎝ weighted sub - sampling rate ⎠
For volunteers,
w8 = w7
The final number of records was 20,832.
Special Surveys Division
51
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
10. Calibration to known population totals
An adjustment was made to the weights in order to make population estimates consistent with
external population counts for persons 15 years and older. The following external control totals
were used:
•
Population totals for each province/census metropolitan area (CMA) stratum, and
•
Population totals by province, sex and the following age groups: 15 to 19, 20 to 24, 25 to
29, 30 to 34, 35 to 39, 40 to 44, 45 to 49, 50 to 54, 55 to 59, 60 to 64, 65 to 69 and 70
and over.
This calibration step was performed merely as a temporary adjustment before identifying outliers.
Once outliers were identified, this calibration step was ignored.
11. Identification and treatment of outliers
The treatment of outliers is a process which diminishes the impact of outlying weighted values.
Outliers were identified for two variables: the total hours volunteered (VD1DHRS) and the total
value of donations (GS1DATOT). Once the outliers were identified, their impact on the total
estimates was diminished by reducing the weight ( w8 ) from Step 9, using a winsorization
technique. The weight of the outlier was reduced such that the adjusted weighted value of the
outlier was equal to the weighted value of the largest non-outlier.
The resulting weight from this step was w9 .
12. Calibration to known population totals
The calibration at this step was performed in the same manner as in Step 10, the only difference
being the weights input into the calibration process. The input to this calibration was the set of
weights, w9 , output from Step 11, after adjusting for outliers. After the calibration was complete,
the outlier detection was performed again to ensure there were no outliers remaining.
The weight, w10 , produced at this step, is the final weight, WTPM, on the Master microdata file
and WTPP on the Public Use Microdata File.
12.2 Weighting for the territorial component
The following steps describe how the weights for the territorial component were calculated.
1. Calculation of initial weights
Because the sample for territorial component of the CSGVP was selected from the Labour Force
Survey (LFS) sample, the initial weight, w1 , was calculated based on design information from the
LFS. The initial weight reflected the inverse of the initial probability of selection.
2. Adjustment of initial weights for non-response
The CSGVP sample can be considered as being comprised of four groups:
1) respondents;
2) units determined to be out-of-scope;
3) non-respondents, resolved to be in-scope; and
4) non-respondents whose in-scope/out-of-scope status is unresolved.
52
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
Each of the 1,831 sample units in the territorial component were assigned a status defined by
these four groups based on the outcome code of the collection application. Since the final
weights of the 1,332 respondents should reflect the entire in-scope population, the weights of the
in-scope respondents should be inflated to account for the non-respondents. The weights should
also be adjusted to account for the fact that the fourth group contains both in-scope and out-ofscope units. Assuming that the proportion of units that are out-of-scope among the unresolved
units is the same as the proportion of out-of-scope among the resolved units, the weights of the
respondents can be adjusted for non-response using the following formula:
w2 gi = w1 gi *
( ∑ w1 gi +
Re sp
∑w
1 gi
NR ,res
∑w
1 gi
Re sp
) ( ∑ w1 gi + ∑ w1 gi + ∑ w1 gi + ∑ w1 gi )
Re sp
NR ,res
OOS
NR ,unres
*
( ∑ w1 gi + ∑ w1 gi + ∑ w1 gi )
Re sp
NR ,res
OOS
where
g represents the level at which the adjustment is performed
w1 gi equals the initial weight of unit i in adjustment group g
∑w
1 gi
equals the sum of the initial weights of all respondents in adjustment group g
Re sp
∑w
1 gi
equals the sum of the initial weights of all resolved non-respondents in adjustment
NR , res
group g
∑w
1 gi
equals the sum of the initial weights of all unresolved non-respondents in adjustment
NR ,unres
group g
∑w
1 gi
equals the sum of the initial weights of all out-of-scope units in adjustment group g
OOS
This adjustment was performed within each stratum, provided that the total number of
respondents plus non-respondents was greater than 30 and the adjustment factor was less than
two. If these conditions did not hold, strata were combined for adjustment purposes. There were
four strata were the sample size warranted a collapsing of strata.
3. Adjustment for sampling one person per household
The weight calculated in Step 2 was multiplied by the number of people in the household 15
years of age or older. In order to avoid problems with outliers and to be consistent with the
weighting procedure for the provincial component, a limit of five was placed on this adjustment.
There were 11 cases where the number of persons in the household aged 15 or older was
greater than five.
w3 = w2 * (number of household members aged 15 + )
4. Calibration to known population totals
The calibration step ensures that the sum of the weights of the respondents is equal to known
population counts. The calibration was performed using age group/sex control totals by territory,
with the three age groups being ages 15 to 24, 25 to 54, and 55 and over. In addition, in Nunavut
the calibration also included a control total for the Inuit population aged 15 and over. The control
totals used were for the October 2004 reference month.
Special Surveys Division
53
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
This calibration step was performed merely as a temporary adjustment before identifying outliers.
Once outliers were identified, this calibration step was ignored.
5. Identification and treatment of outliers
The treatment of outliers is a process which diminishes the impact of outlying weighted values.
Outliers were identified for two variables: the totals hours volunteered (VD1DHRS) and the total
value of donations (GS1DATOT). Once the outliers were identified their impact on the total
estimates was diminished by reducing the weight ( w3 ) , from Step 3, using a winsorization
technique. The weight of the outlier was reduced such that the adjusted weighted value of the
outlier was equal to the weighted value of the largest non-outlier.
The resulting weight from this step was w4 .
6. Calibration to known population totals
The calibration at this step was performed in the same manner as in Step 4, the only difference
being the weights input into the calibration process. The input to this calibration was the set of
weights ( w4 ) , output from Step 5, after adjusting for outliers. After the calibration was complete,
the outlier detection was performed again to make sure that there were no outliers remaining.
The weight, w5 , produced at this step, is the final weight, WTPM, on the Master microdata file and
WTPP on the Public Use Microdata File.
54
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
13.0 Questionnaires
Refer to CSGVP2004_QuestE.pdf for the English questionnaire used for the 2004 Canada Survey of
Giving, Volunteering and Participating (CSGVP).
Special Surveys Division
55
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
14.0 Structure of the files
There are two data files for the 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP):
the main answer file (MAIN.TXT), and the giver file (GS.TXT). To link between the MAIN and GS Master
files use the variable MASTERID and to link between the two Public Use Microdata Files use the variable
PUMFID.
MAIN.TXT
This is the main answer file and contains one record per respondent. All questions except for those on
the GS file are located here. In addition, summary derived variables have been created from the GS file
and placed on the MAIN file.
GS.TXT
This is the “giving” or charitable donation answer file. It contains one or more records for each person
who made a financial donation: one record for each of up to 10 charitable organizations to which the
respondent donated, over the 12 month reference period, in response to a particular solicitation method.
For each of the 13 methods of solicitation itemized in the questionnaire, a donor may therefore have up to
10 records, each containing information regarding the type of organization, as well as the total value of all
donations made to that organization in response to that method of solicitation. In cases where the
respondent donated to more than 10 organizations in response to a given method of solicitation, the total
value of all donations made to the remaining organizations is present on the 10th record as derived
variable GS1D08.
Special Surveys Division
57
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
15.0 Variable naming conventions
The 2004 Canada Survey of Giving, Volunteering and Participating (CSGVP) has adopted a standard
eight character variable naming convention for variables on the microdata files.
Variable name component structure
•
The first two characters are a combination of letters that identify the section of the questionnaire
in which the variable was collected or from which the data used to derive the variable came.
Positions
1 and 2
Questionnaire section name
Positions
1 and 2
Questionnaire section name
FV
Formal Volunteering
GS
Giving Specifics
HV
History of Volunteering
DG
Decisions on Giving
VS
Volunteer Specifics
RG
Reasons for Giving
VD
Volunteer Details
NG
Reasons for Not Giving (more)
MV
Main Volunteer Activities
OG
Other Giving
RV
Reasons for Volunteering
HG
Health in General
GV
Volunteering in General
PA
Participating
ES
Employer Support
ED
Education
NV
Reasons for Not Volunteering (more)
LF
Labour Force Status
IV
Informal Volunteer Activity
SD
Socio-demographics
FG
Financial Giving to Charitable
Organizations
IN
Income
•
The third character of the variable name is an identifier of the “wave” or iteration of a longitudinal
survey. This is always equal to “1” on the 2004 CSGVP.
•
The fourth character of the variable name refers to the variable type.
Position
4
Variable
type
Description
_
Collected
variable
A variable that appeared directly on the questionnaire
C
Coded
variable
A variable coded from one or more collected variables (e.g., National Occupational
Classification – Statistics)
D
Derived
variable
A variable calculated from one or more collected or coded variables, usually calculated
during head office processing (e.g., total hours volunteered)
F
Flag variable
A variable calculated from one or more collected variables (like a derived variable), but
usually calculated by the computer application for later use during the interview (e.g.,
volunteer flag).
G
Grouped
variable
Collected, coded, suppressed or derived variables collapsed into groups (e.g., age
groups)
I
Imputation
flag
A flag indicating whether a particular variable has been imputed (not present on the
Public Use Master File).
Special Surveys Division
59
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
•
The fifth, sixth, seventh and eighth characters identify the variable or the question number from
the questionnaire. In general, the last four positions follow the naming on the questionnaire.
Numbers are used where possible (e.g., Q01 becomes 01). “Mark-all that apply” type questions
use letters for each possible answer category (e.g., Q01 (Mark all that apply) becomes 01A, 01B,
01C, etc.).
Examples of variable names
MV1_02A: Number of hours spent canvassing for the main volunteer organization
MV
Main Volunteer Activities section of the questionnaire
1
2004 CSGVP
_
Collected variable
02
Question number from questionnaire
A
First category in a “Mark all that apply” type question
FV1FVOL: Volunteer flag
FV
Formal Volunteering section of the questionnaire
1
2004 CSGVP
F
Flag
VOL
Variable name
Note: A few important variables do not follow the naming convention (e.g., MASTERID, PUMFID,
PROVCODE, WTPM and WTPP).
60
Special Surveys Division
Canada Survey of Giving, Volunteering and Participating, 2004 – User Guide
16.0 Record layout with univariate frequencies
Refer to CSGVP2004_MAIN_Master_CdBk.pdf for the English record layout with univariate counts for the
Main Master file.
Refer to CSGVP2004_GS_Master_CdBk.pdf for the English record layout with univariate counts for the
Giving (or charitable donation) Master file.
Refer to CSGVP2004_MAIN_PUMF_CdBk.pdf for the English record layout with univariate counts for the
Main Public Use Microdata file.
Refer to CSGVP2004_GS_PUMF_CdBk.pdf for the English record layout with univariate counts for the
Giving (or charitable donation) Public Use Microdata file.
Special Surveys Division
61